Generated by StatCvs v0.2-dev
Pages: << Previous 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 Next >>
Summary Period: 2000-03-31 to 2000-03-23 (Commits 2045-2094 of 12744)
Greatly simplified the new infinite renderer. Also removed
a unix dependency.
3 lines of code changed in:
- Added static compilation support for line renderer and new
infinite renderer.
- Added new infinite 3D renderer for statistics gathering. This
is a theoretical infinite speed renderer with which you can
exactly time what work the engine is doing and what the
maximum theoretical FPS for the engine is.
This renderer doesn't even create a window so it is not
interactive at all. This means that you have to prepare a
bit in order to use it. First start with a normal renderer
and go to some spot that you want to benchmark. Press shift-C
there to record the position (in 'coord'). Then quit and
restart walktest with the following parameters:
walktest level -video=infinite -exec=perf.scr
perf.scr is an example script that will turn on the c-buffer,
restore the camera to the location, render a number of frames
and then quit CS.
- Added -overdraw commandline option for the infinite 3D renderer
so that you have additional testing for the overdraw ratio
per frame. The overdraw ratio is the ratio of drawn pixels
compared to the number of screen pixels. A ratio of one means
no overdraw. A ratio less than one means underdraw (i.e. not
all of the screen is rendered). Note that since the default
clipper in CS is actually slightly less than the screen size
you will get something like 0.98 also even if you have perfect
no overdraw.
Note that overdraw testing SLIGHTLY increases the infinite
3D renderer processing time so this means that performance
might go down a little.
- Added -fastmesh commandline option for the infinite 3D
renderer which simulates a renderer that has hardware accelerated
transforms. In other words, DrawTriangleMesh and DrawPolygonMesh
will not do any transformation from world->screen space anymore
as it happens when -fastmath is not used. Note that this
invalidates the accuracy of the overdraw testing. So you should
probably not combine both options.
- Added argument to 'perftest' command for the number of
performance iterations.
87 lines of code changed in:
Converted to LF format.
158 lines of code changed in:
No longer emits a final trailing command separator (";") when copying
source images in the %.IMAGES target since some shells fail to grok such
an occurence and print a "missing command" error message. In particular,
the DJGPP shell seems to have this problem.
7 lines of code changed in:
A small update to Texinfo documentation from Martin Geisse.
6 lines of code changed in:
Made the following changes to csProgressMeter:
-*- Added a "refresh granularity" setting which allows the client to
configure how often the meter is refreshed. The default is 10, which
means that the meter will refresh every 10%, or at most 10 times.
-*- Added a "tick scale" setting which allows the client to control the
number of units which a single tick mark represents. Default is 2,
which means that the meter is represented by a total of 50 tick marks.
-*- Added a Restart() method which both resets the meter and prints the
initial tick mark ("0%").
-*- Meter is now displayed via iSystem::Printf() rather than CsPrintf()
which means that it is no longer unnecessarily bound to the csEngine
library. An iSystem pointer is passed to the constructor at creation
time.
-*- Documented all public methods with Doc++ style comments.
-*- Now explicitly prevents overflow of its refresh buffer.
131 lines of code changed in:
Mingw makefile process
165 lines of code changed in:
-*- "-gamma=#" option works again (for software renderer, of course).
Alas, other renderers still lack gamma support.
-*- Gamma has now the correct direction. E.g. 2.0 for a brighter image,
0.5 to darken the image, not vice versa as it was before.
wouter@ppp14.dialup.cs.vu.nl told me Quake has a incorrect direction
of gamma too.
12 lines of code changed in:
Update to visibility document. This is the first draft.
19 lines of code changed in:
Image loader fixes. OpenStep compiler work-arounds.
10 lines of code changed in:
- Fixed a few things in MazeD. All polygons are now unnamed by default,
unless you give them a name.
- Made Home/End go to top/bottom of lists (Ctrl+PgUp/Dn works too - like before).
To go horizontally to begin/end use Ctrl+Home/Ctrl+End
15 lines of code changed in:
Changed QInt back (seems working)
0 lines of code changed in:
Worked around OpenStep bug where compiler generates unlinkable code when
an line method calls a second inline method which the compiler has not yet
seen a declaration for. Fixed by reorganizing file textually.
4 lines of code changed in:
Added X11/gpm/slang libs for Unix
4 lines of code changed in:
- Fixed some minor bugs in MazeD
- Fixed the polygon clipper to not give divide overflows anymore
6 lines of code changed in:
Fixed some bugs that popped up when DECLARE_TYPED_VECTOR was changed to delete
it's items. It resulted in some better data management but broke some
assumptions made when writing the code originally.
Added the first three chapters of a tutorial for adding CrystalClear
functionality to Simple.
1 lines of code changed in:
- Fixed a bug in yesterdays bugfix on CompressShadowFrustrums.
- Fixed a bug which caused sectors using STATBSP to work wrong
when the sector has less than 150 polygons.
2 lines of code changed in:
Update.
6 lines of code changed in:
Started new visibility doc.
1 lines of code changed in:
Fixed a warning.
1 lines of code changed in:
Several minor code fixes.
2 lines of code changed in:
Eliminated compilation warning.
1 lines of code changed in:
Once again compiles on NextStep.
7 lines of code changed in:
Eliminated a couple compilation warnings.
2 lines of code changed in:
Eliminated a compilation warning (float -> int).
24 lines of code changed in:
Fixed a bug which prevented lighting to work properly when
using multiple STATBSP levels connected through portals.
This bug was documented actually in the source code :-)
(CompressShadowFrustrums).
41 lines of code changed in:
Fixed a botch.
1 lines of code changed in:
- Fixed unlighted polygon saving in MazeD
- Temporarily replaced QInt(x) with just int(x) so that we can find out
whenever the -recalc problems reported by Jorrit are related to QInt
- Resolved some memory leaks in cssprite.cpp
46 lines of code changed in:
-*- Fixed the WAV loader so that it works correctly for Jorrit on
big-endian Sparc machines.
-*- Replaced non-standard PORT_BYTESEX_BIG_ENDIAN with CS_BIG_ENDIAN.
36 lines of code changed in:
Removed numerous, extremely outdated makefile support files from CS/mk
and subdirectories. These files were generally three to six months out of
date (which is very significant in CS-time :-).
4 lines of code changed in:
- Added csColor+= operator to add colors.
- Fixed bug in loader of new mixmode parameter (for sprites).
MULTIPLY2 was not correctly parsed.
- Added new LIGHTING(yes/no) keyword for 2D sprites to enable
or disable lighting. The COLORS keyword no longer disables
lighting by default. Instead it allows you to define extra
colors to add to lighting (if lighting is enabled).
42 lines of code changed in:
Fixed a typographical error and added more content to authors.txi.
10 lines of code changed in:
*** Phase Three of Eric's Big Documentation Overhaul {tm} ***
-*- Created a custom tool (CS/bin/docconv.pl) for automated conversion of
Texinfo to HTML. Converted files are committed to the repository only
if actual changes have been made to the Texinfo master files, so it is
safe invoke the tool even when no changes have occurred. The script
takes special care to invoke the appropriate CVS commands to add new
files and directories to the repository, and to remove obsolete files.
Furthermore, it correctly handles binary files, such as images, by
committing them to the repository with the "-kb" CVS flag.
-*- Enabled automatic conversion of Crystal Space's Texinfo documentation
to HTML format. Conversion occurs every six hours (or four times
per day).
44 lines of code changed in:
W.C.A. Wijngaards <wouterw@cs.vu.nl> made a small fix to
line renderer.
2 lines of code changed in:
Added support for explicit color setting of the 2D sprite
vertices. This can be used instead of the default lighting
if you want the vertices to have a fixed color.
Use the COLORS keyword in the loader.
21 lines of code changed in:
Implemented support for lighting of 2D sprites.
56 lines of code changed in:
Update.
39 lines of code changed in:
-*- Further work on MazeD. Well, it is already useable: you can build simple
levels in MazeD (from scratch) without having to change almost anything
in the output world file except adding the START(room,x,y,z) statement.
In fact, I already created some simple rooms for testing.
-*- Added normals display in MazeD. By default this is off, the toggle is
operated with 'N' key. Useful to see from which side a polygon is visible
(also useful to see whenever there is a polygon somewhere at all).
-*- Changed default scale and grid spacing to 1 unit instead of 0.25.
-*- Added saving of texture mapping. Now respective PLANE() and TEXTURE()
statements are created (a plane is created only if some texture mapping
parameters are used more than once).
-*- Changed default light attenuation from "realistic" to "linear". The
"realistic" lighting looks too unrealistic :-)
-*- Added vertex compression routine. It is run after loading a room that
contains a static thing (__static__). I'm planning to make this function
separately available somehow (from menu and maybe via hot keys).
-*- Fixed the routine that computes the polygons that make the convex hull of
a room. It used to grab the wrong polygons sometimes.
-*- When a world file is saved, the "precalc_info" and "lm/*" files are
removed since lighting information have to be recalculated anyway.
11 lines of code changed in:
-*- Fixed the halo drawing routine in software renderer. It displayed incorrect
halos for very bright (R/G/B > 1) lights.
29 lines of code changed in:
Updated docs/incoming/worldfmt.tex to conform to the latest
changes related to TEXTURES, 2D sprites, and UVEC/VVEC.
3 lines of code changed in:
- Started some serious optimization on sprites. In OpenGL with
10 sydneys I could only get about 5 FPS. Now I get 13 FPS.
- Optimized sprites a bit more.
- Changed curve drawing using new clipper routine to get
u,v coordinates.
- Continued work on 2D sprites. It is now possible to draw
them. Lighting does not work yet.
159 lines of code changed in:
-*- Modified funConsole so it is now a part of the cscon plugin instead of a
separate plugin that depends on cscon (this is how it is already in MSVC).
-*- Changed cctest so it shows a fullscreen startup console but a smaller
interactive console so you can still see funConsole's border.
-*- Added funConsole to scf.cfg and made it the default console for cctest.
-*- Changed the way cctest handles mouse input. Should be more friendly on
Windows, but I haven't tested it there yet.
-*- cctest no longer automatically captures the mouse. You click on the window
to activate mouse capture and press ESC to release it. Also made the right
mouse button move backwards for better control from the mouse.
-*- Increased of the speed of the rockets. It is no longer possible for the
user to outrun them. :-)
3 lines of code changed in:
Fixed a scoping issue for MSVC and NextStep ports
3 lines of code changed in:
Patched the Texinfo to HTML conversion rule to work on DOS, by moving some
of the texi2html options from the command-line into the 'init' file.
This works around DOS's 128-byte command-line length limitation.
49 lines of code changed in:
- Added csCBufferCube which is a cube of six c-buffers :-)
- Tried to use csCBufferCube for lighting instead of coverage mask tree
cube. I was expecting to get more performance for lighting because
the c-buffer is faster than the coverage mask tree. At least that
was what I thought! But I discovered that the coverage mask tree
implementation does not test for IsFull to stop processing
a tree if it is full! This can make a HUGE difference. So now
I'm not so sure that the c-buffer is really better. This needs
to be reevalulated.
- Removed csQuadtree and csQuadcube usage from the engine since the
quadtree is the worst of all 2D cullers we have now.
- Removed 'quadtree' command from WalkTest.
- Fixed bug in WalkTest movement code on systems that return negative
values for Time() (movement would not be possible on such systems).
- Thanks to Andrew's changes to the clipper I implemented a new
clipper in DrawTriangleMesh. This also solves the bug when sprite
triangles were clipped. Note that this bug ONLY fixes sprites. The
bug will still be present in the other triangle drawers.
This fix fixes Squawk as well I think.
220 lines of code changed in:
-*- Introduced two new keywords in world file (to make them smaller when
generating world files automatically such as from MazeD or map2cs):
UVEC(x,y,z)
VVEC(x,y,z)
They are a replacement for FIRST(x,y,z)/FIRST_LEN(l) and
SECOND(x,y,z)/SECOND_LEN(l). The UVEC() defines the vector of
top texture edge, and the VVEC() defines the vector of left texture
edge. The length of the vector is same value that is usually specified
with FIRST_LEN and SECOND_LEN. Also take into account that the vectors
start from ORIG() point. Thus, the sequence:
ORIG (10, 10, 10)
FIRST (11, 10, 10)
FIRST_LEN (10)
SECOND (10, 9, 10)
SECOND_LEN (5)
can be written now as:
ORIG (10, 10, 10)
UVEC (10, 0, 0)
VVEC (0, -5, 0)
These two keywords are recognized in PLANE(), POLYGON() and polygon
template, BEZIER() and bezier template (although it looks like it is
unused there yet) contexts.
-*- Implemented texture mapping saving in MazeD; about 90% done.
6 lines of code changed in:
-_- Fixed breakage from updated csPixmap
-_- Fixed duplicate default parameters in csMD5 and ccActor
2 lines of code changed in:
added GLIDE_V3 definition
2 lines of code changed in:
added line for glide3 driver
1 lines of code changed in:
Update.
19 lines of code changed in:
Pages: << Previous 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 Next >>
Generated by StatCvs v0.2-dev